Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Areeb Jamal
    @iamareebjamal
    Hence, nginx must accept plain HTTP instead of HTTPS
    So, apache config should be changed to reverse proxy to http nginx, not https
    Norbert Preining
    @norbusan
    Hmm, apache is running on the host, nginx on the vm, apache on the host proxies both https and http to the respective ports of the vm
    Areeb Jamal
    @iamareebjamal
    That's the issue, apache terminates the HTTPS connection and now the request is plain HTTP and sends to 443 port of VM which rejects it. Either apache should not terminate HTTPS or should send it to 80 port of VM only
    Norbert Preining
    @norbusan
    Hmm, how to "not terminate" the https connection?
    I redirected to :80 from apache, can you check now?
    Areeb Jamal
    @iamareebjamal
    Now, it is just redirecting infinitely
    Norbert Preining
    @norbusan
    Yes, becasue ngincx on :80 redirects to voicerepublic.com:443 and back 80 ... umpf ...
    Areeb Jamal
    @iamareebjamal
    Because https also redirects to 80 and 80 in VM redirects to https
    Can you remove the certs from apache config?
    Norbert Preining
    @norbusan
    And then?
    Then it does break, surely
    Areeb Jamal
    @iamareebjamal
    nginx must have certs right?
    Norbert Preining
    @norbusan
    The first connection is to the host's apache server, and the ssl connection is set up there, so the certs must be there.
    ssl cannot be proxied through!!!
    Areeb Jamal
    @iamareebjamal
    If you want apache to handle the certs, then nginx
    config must be changed
    so that it does not redirect on 80
    Norbert Preining
    @norbusan
    Why was it working on the other VM,then?
    Areeb Jamal
    @iamareebjamal
    Maybe the nginx was not behind apache there?
    Cloudflare proxies HTTPS requests as well, right? First they terminate HTTPS at their server, and then recreate HTTPS to the destination server. In this case, apache does not recreate HTTPS connection to nginx and just sends plain HTTP to port 443
    Norbert Preining
    @norbusan
    Haaa, ...
    working now ....
    stupid me ....
    copy and paste is not a good idea ...
    http://....:443 simply does not work :-D
    Areeb Jamal
    @iamareebjamal
    Must be https
    admin site is broken now though
    Reason: Error during SSL Handshake with remote server
    Norbert Preining
    @norbusan
    Reload, and do you have admin.voicerepublic.com in your hosts file?
    Here it is working (though logging in seems not to work, or I don't have an account)
    Areeb Jamal
    @iamareebjamal
    Yes, working :+1:
    Mario Behling
    @mariobehling
    Great team work!
    @norbusan Did you sleep at all this night or going to bed now?
    Norbert Preining
    @norbusan
    @mariobehling no, I usually get up at 4am ....
    @mariobehling I still cannot log in with the credentials you sent me. Could you try to log in with yours. You need to add the following stuff to the /etc/hosts file temporarily:
    95.216.25.247   www.voicerepublic.com
95.216.25.247   admin.voicerepublic.com
95.216.25.247   voicerepublic.com
    Norbert Preining
    @norbusan
    Maybe also faye.voicerepublic.com ?
    Mario Behling
    @mariobehling
    Screenshot from 2020-01-23 22-48-27.png
    Norbert Preining
    @norbusan
    Strange ...
    Norbert Preining
    @norbusan

    @mariobehling Proxy Error is fixed, the nginx in the VM uses TLSv1 which was disabled on my apache. Enabling for this connection made it work, connection to admin.voicerepublic.com is now working.

    But after logging, I get a 404 error - ata least progress ;-)

    Norbert Preining
    @norbusan
    The reason is that there seems to be some problem with the postgres database ... 
    E, [2020-01-24T05:45:03.016445 #3488] ERROR -- : PG::InternalError: ERROR:  missing chunk number 0 for toast value 1334328 in pg_toast_2619
: SELECT "talks".* FROM "talks" WHERE (ends_at > '2020-01-24 00:45:03.011660' AND starts_at < '2020-01-24 10:45:03.011933')
    Norbert Preining
    @norbusan
    Anyone knows where the passwords for the backend postgres user is located?
    Norbert Preining
    @norbusan
    Forget that, user based

    Ok, the postgres table is broken

    SELECT * FROM talks LIMIT 1 offset $i

    gives error for each and every row. That means, I need a new pg dump ...

    Mario Behling
    @mariobehling

    @norbusan Here is the feedback from @telmich. He can provide the DB, but..

    I can do that, but there should already be tons of postgresql backups on the VM
    And to get the data, you can simply run pg_dump on the running VM

    Norbert Preining
    @norbusan
    @mariobehling Yes, there are many dumps, but aren't there any newer data since then? I can restore from the latest one for sure!
    Norbert Preining
    @norbusan
    I have now pulled a new DB dump from the current VR server and played it into my VM. Now the admin backend is working.
    @mariobehling please try out the admin backend, and if possible other features, too, so that we see whether the functionality is fully back up!
    Norbert Preining
    @norbusan
    Frontend seems to work, too, at least I could listen to a few talks.